Pricing Docs Playground Use cases Verified Company Get API Key Try Playground
Deterministic language data from the edge

Grounded vocabulary for AI agents. No hallucinations. No model drift.

Word Orb gives your agent verified definitions, IPA pronunciation, translations in 47 languages, and structured lessons — as deterministic JSON from the Cloudflare edge. Same input, same output, every time.

Try It Live Read Docs
Pricing Get free API key Use cases Education
162,253 verified words. 601,143 translations. 45,916 assessment items. Every response is deterministic JSON — no model inference, no token cost, no drift. Free tier: 10,000 calls/day.
162,253
Words
601,143
Translations
45,916
Assessment Items
30,526
Graph Links

Four layers of verified data, one API key

Word Orb

162,253 words · 47 dictionary languages · 601,143 translations

Definitions, IPA pronunciation, etymology, native-script translations, and vocabulary context that developers can trust in production.

  • One API call returns definition, translations, pronunciation, etymology, and audio URL
  • Age-calibrated definitions: child, teen, adult, elder
  • Batch lookup up to 50 words per request

Lesson Orb

Daily lesson infrastructure · 19 lesson languages · 4 tracks · 10 supported archetypes

Structured lesson context around the shared daily experience: five phases, four tracks, and the pedagogical layer behind Kelly.

  • 5 lesson phases: hook, story, wonder, action, wisdom
  • 4 tracks: learn, grow, teach, trivia
  • Every lesson links to vocabulary and assessments through the knowledge graph

Assessment Layer

45,916 assessment items · 6 question types

Assessment objects aligned to vocabulary and lesson structure, so teams can test exactly what they teach.

  • Questions test exactly what the lesson taught — graph-aligned, not random
  • Random, by-day, and batch modes with difficulty filtering
  • Each question includes the answer and an explanation

Knowledge Graph

30,526 connections · words ↔ lessons ↔ assessments

The graph is the glue. It lets teams trace a word into the lessons that teach it and the assessment items that verify it.

  • Trace any word to find all lessons where it appears
  • Find all vocabulary words in any lesson
  • Build complete teach-then-test cycles from three API calls

Three API calls, zero hallucinations

curl JavaScript Python
# 1. Look up the word curl https://wordorb.ai/api/word/courage \ -H "Authorization: Bearer wo_YOUR_KEY" # 2. Get a structured lesson curl "https://wordorb.ai/api/lesson?day=47&track=learn" \ -H "Authorization: Bearer wo_YOUR_KEY" # 3. Assess comprehension curl "https://wordorb.ai/api/quiz?day=47&track=learn" \ -H "Authorization: Bearer wo_YOUR_KEY"
const API = 'https://wordorb.ai/api'; const h = { 'Authorization': 'Bearer wo_YOUR_KEY' }; // 1. Look up the word const word = await fetch(`${API}/word/courage`, {headers: h}) .then(r => r.json()); // 2. Get a structured lesson const lesson = await fetch(`${API}/lesson?day=47&track=learn`, {headers: h}) .then(r => r.json()); // 3. Assess comprehension const quiz = await fetch(`${API}/quiz?day=47&track=learn`, {headers: h}) .then(r => r.json());
import requests API = "https://wordorb.ai/api" h = {"Authorization": "Bearer wo_YOUR_KEY"} # 1. Look up the word word = requests.get(f"{API}/word/courage", headers=h).json() # 2. Get a structured lesson lesson = requests.get(f"{API}/lesson?day=47&track=learn", headers=h).json() # 3. Assess comprehension quiz = requests.get(f"{API}/quiz?day=47&track=learn", headers=h).json()

Why teams switch to verified data

LLMs improvise definitions. Word Orb doesn't. Here's why that matters for production systems.

Deterministic

Same word, same JSON, every time. No temperature, no sampling variance, no model version roulette. Your tests pass tomorrow the same way they pass today.

Auditable

Every definition, translation, and assessment is human-verified and traceable through the knowledge graph. SOC 2 reviewers can see exactly where the data came from.

Edge-fast

Served from 300+ Cloudflare locations. Sub-100ms globally. No cold starts, no token queues. Your agent gets the answer before the user finishes typing.

Works with every agent framework

One config line connects Word Orb to your agent. MCP standard — works with Claude, ChatGPT, Cursor, Gemini, and any MCP-compatible client.

MCP Config
{"mcpServers":{"word-orb":{"url":"https://wordorb.ai/mcp"}}}

AI Agents

Ground your agent's vocabulary in verified data. One MCP tool call replaces hallucinated definitions with deterministic JSON.

Learning Products

365 days of structured curriculum with vocabulary, assessments, and knowledge graph links. Education guide →

Robotics

Pronunciation audio, age-calibrated definitions, and structured JSON for embodied systems. Robotics guide →

Start building in 60 seconds

Free tier: 10,000 API calls per day. All endpoints. No credit card. Ship to production on the free plan.

Try the playground first Read the docs
Latest from the blog
Why Your AI Agent Shouldn't Define Words
npm install wordorb pip install wordorb